System and method for synthesizing low-light images
AssigneeSamsung Electronics Co., Ltd.
InventorsAbdullah Abuolaim, Abhijith Punnappurath, Abdelrahman Abdelhamed, Michael Scott Brown, Aleksai Levinshtein
Life of this patent
Estimating the term…
Abstract
A method of generating a low-light image is provided. The method includes receiving a raw image, removing an amount of first illumination from the raw image, applying a low exposure condition to the raw image having the amount of first illumination removed, and applying an amount of low-light illumination to the raw image having the applied low exposure condition.
Claims17
A method of generating a low-light image, comprising:
- obtaining a raw image captured by a sensor;
- removing an amount of first illumination from the raw image;
- applying a low exposure condition to the raw image having the amount of first illumination removed;
- obtaining a low-light image by applying an amount of second illumination less than the amount of first illumination to the raw image having the applied low exposure condition;
- obtaining a noisy image by applying an amount of noise to the low-light image having the applied amount of second illumination;
- and
- training a neural network model for denoising an image based on the low-light image and the noisy image.
A system for generating a low-light image, comprising:
- a memory storing instructions;
- and
- a processor configured to execute the instructions to:
- obtain a raw image captured by a sensor;
- remove an amount of first illumination from the raw image;
- apply a low exposure condition to the raw image having the amount of first illumination removed;
- obtain a low-light image by applying an amount of second illumination less than the amount of first illumination to the raw image having the applied low exposure condition;
- obtain a noisy image by applying an amount of noise to the low-light image having the applied amount of second illumination;
- and
- train a neural network model for denoising an image based on the low-light image and the noisy image.
A non-transitory computer-readable storage medium configured to store instructions that, when executed, cause at least one processor to:
- obtain a raw image captured by a sensor;
- remove an amount of first illumination from the raw image;
- apply a low exposure condition to the raw image having the amount of first illumination removed;
- obtain a low-light image by applying an amount of second illumination less than the amount of first illumination to the raw image having the applied low exposure condition;
- obtain a noisy image by applying an amount of noise to the low-light image having the applied amount of second illumination;
- and
- train a neural network model for denoising an image based on the low-light image and the noisy image.
Dependencies are inferred from the claim text.